JP2016223795A - Floor surface state detector and floor surface state detecting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a liquid on a floor surface in a non-contact manner.SOLUTION: A distance image acquisition part 10 acquires a first distance image and a second distance image. The first distance image uses a light with such a wavelength that makes the light reflect from a liquid and a floor surface. The second distance image uses a light with such a wavelength that makes the light transmit through a liquid and reflect from a floor surface. Next, a difference extraction part 11 extracts the difference between the first and second distance images. A liquid detection unit 13 thereafter conducts matching with the difference, using shape data on a large number of liquids stored in a storage part 12, and presumably determines that a liquid exists on the floor surface if the value with the largest correlation of the correlations between the shape data on the difference and the shape data on the various different types of liquids is not smaller than a predetermined threshold value.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a floor surface state detection apparatus and a floor surface state detection method that detect liquid existing on a floor surface in a non-contact manner. In particular, the present invention relates to an apparatus for detecting a liquid using a distance image.

  When there is a liquid such as water on the floor, there is a risk that a person will slip and fall with the liquid. For example, in a medical facility or a nursing facility, a patient may vomit or vomit, and liquid may be present on the floor surface. In addition, the floor near the entrance of a building may get wet with water on a rainy day, and the floor near the water in a bathroom, toilet, bathroom, etc. may get wet with spilled water. In such a case, there is a risk of falling as well. Further, when driving an automobile or the like, a puddle on the road surface causes an accident. In order to prevent such accidents, a technique for early detection of liquid on the floor surface is required.

  As a method for detecting the state of the floor surface in a non-contact manner, a method is widely known in which a reflection characteristic is measured by irradiating electromagnetic waves such as infrared rays, and a determination is made from the difference in the reflection characteristic. In addition, there are methods described in Patent Documents 1 to 4.

  Patent Document 1 describes a method of irradiating a road surface with electromagnetic waves at a plurality of frequencies, measuring the frequency dependence of reflection intensity, and determining the road surface state based on the characteristics.

  Patent Document 2 describes that a puddle shape on a road surface is detected by measuring an uneven shape of a road surface by measuring a distance using electromagnetic waves or sound waves, and detecting rainfall by a rain sensor.

  Patent Document 3 describes that a far-infrared image is acquired by changing the temperature of the substrate of the image sensor, and a puddle on the road surface or freezing of the road surface is detected based on a difference in the far-infrared image due to a difference in substrate temperature. Yes.

  Patent Document 4 discloses a method of photographing a road surface with a camera, analyzing an image to create a luminance histogram, and determining a road surface state based on the histogram.

  In addition, distance image sensors that measure an object three-dimensionally are known (for example, Patent Documents 5 and 6). A distance image has distance information in each pixel of image data, and a TOF (Time of Flight) method is generally used as a distance measurement method. TOF is a method for obtaining a distance according to the time until a radiated electromagnetic wave or sound wave is reflected by an object and returns. In addition, it is conventionally known to acquire a distance image by a distance measurement method such as a triangular distance measurement method or a pattern irradiation method.

JP 2008-180623 A JP 2010-257307 A JP 2008-236062 A Patent No. 5190204 JP 2009-124398 A JP 2014-160014 A

  The conventional method for detecting the floor surface state is to determine the floor surface state by measuring the reflection intensity, but this method can only detect that an object is actually present on the floor surface. It was difficult to determine whether or not the liquid was liquid. Moreover, in the method of patent document 2, when a puddle exists by factors other than rainfall, or when water exists on the plane without an unevenness | corrugation, it cannot detect.

  Therefore, an object of the present invention is to realize a floor surface state detection device and a floor surface state detection method capable of detecting liquid on the floor surface.

  According to the present invention, in a floor surface state detection device for detecting a liquid existing on a floor surface, the floor surface and the liquid are irradiated with a first electromagnetic wave or a sound wave having a frequency reflected to the liquid and the floor surface. A first distance image acquisition unit that acquires a distance image; and irradiates the floor surface and the liquid with a second electromagnetic wave or a sound wave having a frequency that is transmitted through the liquid and reflected from the floor surface. The difference between the second distance image acquisition unit that acquires the first distance image acquired by the first distance measurement unit, and the second distance image acquired by the second distance measurement unit is extracted. It is a floor surface state detection apparatus characterized by having a difference extraction part and the liquid detection part which detects a liquid from the difference extracted by the difference extraction part.

  Further, the present invention provides a floor surface state detection device for detecting a liquid existing on a floor surface, irradiating the floor surface and the liquid with a first electromagnetic wave or a sound wave having a frequency reflected to the liquid and the floor surface, A first distance image acquisition unit that acquires a distance image of 1; a third electromagnetic wave or a sound wave having a frequency that is absorbed by the liquid and reflected by the floor surface; A difference between the third distance image acquisition unit that acquires the distance image, the first distance image acquired by the first distance measurement unit, and the third distance image acquired by the third distance measurement unit is obtained. It is a floor surface state detection apparatus characterized by having a difference extraction part to extract and a liquid detection part which detects a liquid from the difference extracted by the difference extraction part.

  The liquid to be detected in the present invention is arbitrary as long as it is a liquid object. For example, water, alcohol, or a solution using these as a solvent. The liquid may be a mixture of gas or solid, and may be a sol or gel. Further, the liquid may be frozen or semi-thawed.

  Further, the floor surface in the present invention includes not only the surface of the floor of the building but also the road surface such as a road and the ground. Further, the floor may be flat or uneven.

  Acquisition of the first to third distance images may use electromagnetic waves or may use sound waves. The distance measuring method may be any method as long as it uses electromagnetic wave or sound wave irradiation. For example, a time-of-flight (TOF) method, a triangulation method, a pattern irradiation method, or the like can be used.

  The road surface state detection apparatus of the present invention may further include a storage unit that stores and holds a plurality of liquid feature patterns, and the liquid detection unit stores the difference extracted by the difference extraction unit and the storage unit. The characteristics of the liquid (the shape of the liquid and the characteristics related thereto) may be estimated by comparing with a plurality of characteristic patterns of the liquid. The feature pattern is a pattern regarding the shape of the liquid and features related thereto, such as a liquid shape pattern and a point cloud pattern. By performing matching using the feature pattern in this way, the liquid shape can be detected with higher accuracy.

  Moreover, you may provide further the contact detection means which dispatches to the position of the liquid detected by the liquid detection part, contacts a liquid, and detects a liquid. By providing the means for directly detecting the liquid in this way, the position and range of the liquid can be detected in more detail.

  Another aspect of the present invention is a floor surface state detection method for detecting a liquid existing on a floor surface, irradiating the floor surface and the liquid with a first electromagnetic wave or a sound wave having a frequency reflected to the liquid and the floor surface, A distance image of 1 is acquired, a second electromagnetic wave or a sound wave having a frequency that is transmitted through the liquid and reflected from the floor surface is irradiated to the floor surface and the liquid, and a second distance image is acquired. The floor surface state detection method is characterized in that the difference between the distance image and the second distance image is extracted and the liquid is detected from the difference.

  According to another aspect of the present invention, in the floor surface state detection method for detecting a liquid existing on the floor surface, the floor surface and the liquid are irradiated with the first electromagnetic wave or the sound wave having a frequency reflected on the liquid and the floor surface, A first distance image is acquired, a third electromagnetic wave or sound wave having a frequency that is absorbed by the liquid and reflected from the floor surface is irradiated to the floor surface and the liquid, a third distance image is acquired, The floor surface state detection method is characterized in that the difference between the first distance image and the third distance image is extracted and the liquid is detected from the difference.

  According to the present invention, the floor surface and the liquid can be distinguished and detected, and the liquid present on the floor surface can be detected.

The figure which showed the structure of the floor-surface state detection apparatus of Example 1. FIG. 3 is a flowchart illustrating the operation of the floor surface state detection device according to the first embodiment. FIG. 3 is a diagram schematically illustrating the principle of liquid detection on the floor surface in the first embodiment. The figure which showed the structure of the floor-surface state detection apparatus of Example 2. FIG. The figure which showed typically the principle of the liquid detection on the floor surface in Example 2. FIG. The figure which showed the structure of the floor-surface state detection apparatus of Example 3. FIG. FIG. 3 is a diagram schematically illustrating the principle of liquid detection on the floor surface in the first embodiment. The figure which showed typically the principle of the liquid detection on the floor surface in Example 2. FIG. The figure which showed the example of the shape of a liquid.

  Specific examples of the present invention will be described below, but the present invention is not limited to the examples.

  FIG. 1 is a diagram illustrating a configuration of the floor surface state detection device according to the first embodiment. The floor surface state detection device according to the first embodiment is a device that detects liquid existing on the floor surface in a non-contact manner. As illustrated in FIG. 1, the distance image acquisition unit 10, the difference extraction unit 11, and the storage unit 12. And the liquid detector 13. The liquid to be detected may be arbitrary. For example, water, alcohol, oil, or those using them as a solvent. Further, the liquid is not limited to a pure liquid, and may be a liquid in which a gas or a solid is mixed, or a sol or gel. Further, the liquid may be frozen or semi-thawed. Further, the floor surface not only means the floor surface of the building but also includes the road surface and the ground.

  The distance image acquisition unit 10 includes two distance image sensors 10A and 10B. The range image sensors 10A and 10B acquire range images by a TOF (Time of Flight) method using electromagnetic waves. The TOF method is a method of measuring the distance based on the time until the light irradiated to the object is reflected and returned. In addition, the distance image is one in which each pixel of the two-dimensional image has distance information. Hereinafter, the distance image acquired by the distance image sensor 10A is referred to as a first distance image, and the distance image acquired by the distance image sensor 10B is referred to as a second distance image.

  The distance image sensors 10A and 10B have a light emitting element, a light receiving element, and a distance calculation unit. The light emitting element is a light emitting diode, a semiconductor laser, or the like, and emits light having a predetermined wavelength whose intensity is modulated. The light receiving element is a two-dimensional image sensor such as a CCD or a CMOS. The distance calculation unit calculates a distance R to the object for each pixel of the two-dimensional image sensor and outputs distance image data. Specifically, the phase difference φ between the modulation phase of the light emitted from the light emitting element and the modulation phase of the light received by each pixel of the two-dimensional image sensor is calculated, and up to the object by R = cφ / (4πf) The distance R is calculated. Here, c is the speed of light and f is the modulation frequency. The maximum measurable distance is c / (2f). The distance calculation unit is realized, for example, by signal processing using an IC (ASIC) specialized for the distance calculation or a computer program.

  The distance image sensors 10A and 10B irradiate light having high directivity such as a laser, scan this two-dimensionally or three-dimensionally, and receive reflected light by a light receiving element to obtain a distance image. There may be.

  Here, the distance image sensor 10A and the distance image sensor 10B have different light emission wavelengths as follows.

  The wavelength of light emitted from the light emitting element of the distance image sensor 10A (hereinafter referred to as the first wavelength) is a wavelength that reflects the liquid and the floor surface. The reflectance with respect to the liquid and the floor surface may be in a range where a distance image by reflection on the liquid surface and a distance image by reflection on the floor surface can be acquired, and it depends on the light receiving performance of the distance image sensor 10A.

  The wavelength of light emitted from the light emitting element of the distance image sensor 10B (hereinafter, the second wavelength) is a wavelength that transmits the liquid and reflects the floor surface. The transmittance with respect to the liquid may be such that a distance image due to reflection on the liquid surface cannot be acquired, and the reflectance with respect to the floor surface may be within a range in which the distance image can be acquired with reflection on the floor surface. It depends on the light receiving performance of the distance image sensor 10B.

  It is desirable to improve the liquid detection accuracy by selecting the light receiving performance of the distance image sensors 10A and 10B so that the difference between the first distance image and the second distance image becomes clearer.

  If the liquid absorption rate is high and it is difficult to acquire the second distance image, the liquid may be detected by the method of Example 2 described later.

  In addition, as long as it has the above-described reflection / transmission characteristics, the object to be irradiated to acquire the first and second distance images is not limited to light but may be electromagnetic waves having an arbitrary frequency (wavelength).

  For example, when the liquid to be detected is water, a reflection band exists in the visible light band, the infrared band, and the like, and a transmission band exists in the infrared band, the millimeter wave band, and the like. The wavelength may be selected as the first and second wavelengths.

  Further, if the first distance image and the second distance image can be acquired, the distance image acquisition unit 10 may not be configured to include the two distance image sensors 10A and 10B. For example, the following configuration may be used. Good.

  Using a distance image sensor equipped with a frequency variable light emitting element, the first distance image and the second distance image can be acquired by changing the light emission wavelength of the light emitting element to the first wavelength and the second wavelength. May be. Alternatively, a first distance image and a second distance image may be obtained by using a distance image sensor equipped with two light emitting elements having different emission wavelengths and alternately switching the light emitting elements to be driven. Also, a light emitting element that emits broad light such as white light is used, two light receiving elements are used, a filter that transmits the first wavelength is provided in front of one light receiving element, and the front of the other light receiving element is provided. By providing a filter that transmits the second wavelength, the first and second distance images may be acquired.

  The difference extraction unit 11 and the liquid detection unit 13 detect liquid from the data of the first distance image and the second distance image from the distance image sensors 10A and 10B. This is realized by signal processing by a computer program, for example. The specific operation will be described later.

  The storage unit 12 holds a lot of liquid shape data. The storage unit 12 is a storage device such as a hard disk or a flash memory. The shape data of the liquid is shape data when the surface shape of the liquid is changed variously by using the floor surface shape, the friction coefficient of the floor surface, the surface tension of the liquid, gravity and the like as parameters.

  Next, operation | movement of the floor-surface state detection apparatus of Example 1 is demonstrated according to the flowchart of FIG.

[Step S1]
First, the distance image acquisition unit 10 acquires a first distance image and a second distance image. The first distance image is acquired using light of a first wavelength that is a wavelength reflected with respect to the liquid and the floor surface. The second distance image is acquired using light having a second wavelength that is a wavelength that transmits the liquid and reflects the floor surface. One of the first distance image and the second distance image may be acquired and then the other may be acquired. However, the first wavelength light and the second wavelength light irradiated for acquiring the distance image interfere with each other. If not, the first distance image and the second distance image may be simultaneously acquired by simultaneously irradiating light.

[Step S2]
Next, the difference extraction unit 11 extracts the difference between the first distance image and the second distance image. Here, the difference is a portion that exists in one distance image but does not exist in the other distance image. That is, a portion that exists only in the first distance image and does not exist in the second distance image, and a portion that exists only in the second distance image and does not exist in the first distance image are extracted.

[Step S3]
Next, in the liquid detection unit 13, if the difference between the first distance image and the second distance image is not extracted in step S2, the process proceeds to step S4. If the difference is extracted, the process proceeds to step S5. Perform forward processing.

  Here, the presence / absence of the difference corresponds to the presence / absence of the liquid. When the difference exists, the liquid exists on the floor surface, and when the difference does not exist, the liquid does not exist on the floor surface. The principle will be described with reference to FIG.

  Consider a case where the liquid 200 exists on the flat floor surface 100 as shown in FIG. The light emitted from the distance image sensor 10A is the first wavelength reflected to the liquid and the floor surface. The light emitted from the distance image sensor 10A is reflected by the liquid surface and the floor and returns to the distance image sensor 10A. Therefore, as shown in FIG. 3B, the first distance image acquired by the distance image sensor 10A is acquired by integrating the surface shape of the floor and the surface shape of the liquid. In FIG. 3 (b), the portion where the distance image can be acquired on the liquid surface and the floor surface is indicated by a solid line, and the portion which is not obtained is indicated by a dotted line. The same applies to FIG. 3C described later. Thus, the surface shape of the liquid cannot be distinguished from the surface shape of the floor surface only by the first distance image.

  On the other hand, the light emitted from the distance image sensor 10B is the second wavelength that is transmitted to the liquid and reflected to the floor surface. Therefore, at the position where the liquid exists, the light of the second wavelength is transmitted and reaches the floor surface, is reflected by the floor surface, and returns to the distance image sensor 10B. At a position where no liquid is present, the light is reflected by the floor and returns to the distance image sensor 10B. Therefore, as shown in FIG. 3C, the second distance image acquired by the distance image sensor 10B acquires the surface shape of the floor surface not only at the position where the liquid does not exist but also at the position where the liquid exists. The surface shape of the liquid is not acquired.

  Comparing FIG. 3 (b) and FIG. 3 (c), FIG. 3 (b) differs from FIG. 3 (c) in that the surface shape of the liquid is acquired. In FIG. It differs from the figure (b) in that the surface shape of the floor surface at the position is acquired. Therefore, when the difference between the first distance image and the second distance image is extracted, the surface shape of the liquid and the surface shape of the floor surface of the area where the liquid exists are extracted as shown in FIG. That is, the shape of the liquid on the floor surface is extracted. In FIG. 3D, the difference between FIG. 3B and FIG. 3C is indicated by a solid line, and the other portions are indicated by dotted lines. Further, as shown in FIG. 3D, since the shape of the floor surface at the position where the liquid exists is also known, the height of the liquid from the floor surface can also be detected. FIG. 3 shows the case where the floor surface 100 is flat. Similarly, when the floor surface 101 is uneven as shown in FIG. 7 and the liquid 201 exists in the recess, the surface shape of the liquid and the liquid The surface shape of the floor surface in the existing area can be extracted. That is, the liquid shape can be extracted regardless of the shape of the floor surface.

  Further, as shown in FIGS. 3B and 3C, the portion that exists only in the first distance image and does not exist in the second distance image corresponds to the surface shape of the liquid, and only in the second distance image. The portion that exists and does not exist in the first distance image corresponds to the shape of the floor surface of the region where the liquid exists. Therefore, if it is sufficient to know the surface shape of the liquid, in the difference extraction in step S2, only the portion that exists only in the first distance image and does not exist in the second distance image is extracted, and the second distance is extracted. A portion that exists only in the image and does not exist in the first distance image may not be extracted. For example, when the shape of the floor surface is known.

  When no liquid is present on the floor surface, only the surface shape of the floor surface is acquired for both the first and second distance images. Therefore, there is no difference between the first distance image and the second distance image, and the difference between the first distance image and the second distance image is not extracted.

  As described above, by extracting the difference between the first distance image and the second distance image, it is possible to detect whether or not the liquid exists on the floor surface, and from the difference, the approximate shape of the liquid , Position, range, quantity, etc. can be detected.

[Step S4]
If the difference between the first distance image and the second distance image is not extracted in step S3, it means that there is no liquid on the floor surface, so the liquid detection unit 13 determines that.

[Step S5]
When the difference between the first distance image and the second distance image is extracted in step S3, the liquid exists on the floor surface.

  Therefore, the liquid detection unit 13 performs matching with the difference using a large number of liquid shape data stored in the storage unit 12. As the matching method, various conventionally known methods can be used. For example, the similarity between the difference shape data and the liquid shape data is evaluated by taking a correlation between the difference shape data and the liquid shape data. Machine learning can also be used.

  If there is no liquid shape data that matches the difference, there is a possibility that a non-liquid shape has been detected or a false detection. For example, among the correlations between the difference shape data and the shape data of various liquids, when the largest correlation value is smaller than a predetermined threshold, or the average of the correlation between the difference shape data and the shape data of various liquids This is the case when the value is smaller than a predetermined threshold value. In such a case, the process may proceed to step S4 to determine that there is no liquid on the floor surface.

  As for the number of data of the liquid shape data, the larger the number, the better the detection accuracy of the liquid shape. On the other hand, the larger the number, the longer the processing time for matching. To do.

  Further, the matching may be performed using not only the shape data but also feature data related to the shape. For example, matching may be performed using point cloud (point cloud) data, and the shape of the liquid or features related thereto may be detected.

[Step S6]
By matching in step S5, the shape closest to the difference shape data among the many shape data stored in the storage unit 12 is estimated as the actual shape of the liquid. For example, the liquid shape data having the highest correlation with the difference shape data is estimated as the actual liquid shape. From the determined liquid shape and the first and second distance images, the position, range, shape, amount, etc. of the liquid on the floor surface can be specified. Of course, it may be determined only whether or not liquid is present. For example, among the correlations between the difference shape data and the shape data of various liquids, the average of the correlation between the difference shape data and the various liquid shape data is the case where the value with the largest correlation is a predetermined threshold value or more If the value is equal to or greater than a predetermined threshold value, it is determined that liquid is present on the floor surface.

  If the resolution of the first and second distance images is sufficiently high, the shape of the liquid can be roughly specified from the difference between the first distance image and the second distance image. If it is not necessary to specify, the step S5 may be omitted, and the position, range, shape, amount, etc. of the liquid may be specified directly from the first and second distance images and their differences.

  Further, since the surface shape of the end of the liquid is known by the estimation of the liquid shape in step S6, the contact angle of the liquid on the floor surface is known. FIG. 9 is a diagram showing a difference in liquid shape due to a difference in contact angle. 9A shows a case where the contact angle is large and 90 ° or less (in the case of immersion wetting), FIG. 9B shows a case where the contact angle is larger than 90 ° (in the case of adhesion wetting), and FIG. Is the case where the contact angle is 0 ° (in the case of extended wetting). In the case of FIG. 9A, the surface shape of the liquid is a lens shape. In the case of FIG. 9B, the liquid has a collapsed spherical shape. In the case of FIG. 9C, the surface shape of the liquid is substantially flat except for the end portions.

  The contact angle is determined by the interfacial tension Fr of the floor surface, the surface tension Fs of the liquid, and the interfacial tension Fi between the liquid and the floor surface, which are determined by the surface characteristics of the floor surface and the liquid. Therefore, if the type of liquid is known or can be estimated, it is possible to estimate the state of the floor surface (floor surface material, water repellency, hydrophilicity, etc.) from the contact angle. On the contrary, if the surface characteristics of the floor surface are measured and stored in advance, it can be used for estimating the type of liquid, and can also be used for estimating the shape of the liquid more accurately.

  In the liquid shape determination by matching with shape data, point cloud data, or the like in steps S5 and S6, machine learning using these data as teacher data may be performed. That is, the characteristics of the liquid may be automatically extracted from the teacher data to determine whether it is a liquid, and if it is a liquid, the shape may be determined. There are various methods for machine learning. One of them is an analysis method using a multi-layer neural network called deep learning (deep learning). By using deep learning, the liquid shape can be determined with very high accuracy.

  As described above, according to the floor surface state detection device of the first embodiment, the floor surface and the liquid can be distinguished and detected, and the liquid existing on the floor surface can be detected in a non-contact manner.

  Since the liquid on the floor surface can be detected at an early stage by the floor surface state detection apparatus of the first embodiment, various measures for preventing the overturn can be taken.

  For example, a warning display is projected on the floor around the liquid, the liquid is colored to make it easier to understand visually, a warning sound is emitted around the liquid, and the presence of the liquid on the floor is reported to a third party. It is possible to prevent people from getting close to the liquid by taking measures such as informing them, patroling poles and fences.

  Also, take measures such as spraying absorbent, desiccant, and anti-slip on the liquid, laying mats etc. on the liquid area, blowing air to dry the liquid, dispatching a cleaning robot to remove the liquid, etc. Thus, even if a person enters the place where the liquid existed, it can be prevented from falling down.

  In addition, if a person is already falling, even if the person falls down, such as by deploying a mat or airbag on the floor or deploying an airbag that was previously worn by the person, And so on.

  Moreover, a risk determination means may be provided, and the above various fall prevention measures may be taken based on the risk evaluation. The risk judgment means considers, for example, information on the liquid (position, range, shape, depth, etc. of the liquid) detected by the floor condition detection device and information on the person around the liquid (position of the person, walking speed, etc.) And determine the risk of a person falling.

  The detection of the liquid by the floor surface state detection apparatus according to the first embodiment can be used in addition to avoiding a human fall risk. For example, it can be used to prevent an accident caused by a puddle in driving a car or the like.

  FIG. 4 is a diagram illustrating the configuration of the floor surface state detection device according to the second embodiment. The floor surface state detection apparatus according to the second embodiment uses the distance image acquisition unit 10 according to the first embodiment as a distance image acquisition unit 20, and the distance image acquisition unit 20 includes a distance image sensor included in the distance image acquisition unit 10. Of 10A and B, only the distance image sensor 10B is replaced with a distance image sensor 10C. Other configurations are the same as those of the floor surface state detection apparatus of the first embodiment.

  The distance image sensor 10C obtains a distance image by a TOF (Time of Flight) method using electromagnetic waves in the same manner as the distance image sensors 10A and 10B. The distance image sensor 10C is a light emitting element, a light receiving element, and a distance calculation. Has a part. Hereinafter, the distance image acquired by the distance image sensor 10C is referred to as a third distance image.

  The wavelength of light emitted from the light emitting element of the distance image sensor 10C (hereinafter, the third wavelength) is a wavelength that is absorbed by the liquid and reflected by the floor surface. Absorption with respect to the liquid means that the light of the third wavelength emitted from the distance image sensor 10C returns to the distance image sensor 10C due to less reflection on the liquid surface and absorption of light by the liquid. And the distance image cannot be acquired (the distance is determined to be infinite), and the reflectance with respect to the floor surface may be within a range in which the distance image can be acquired by reflection on the floor surface. It depends on the light receiving performance of the distance image sensor 10C.

  It is desirable to improve the liquid detection accuracy by selecting the light receiving performance of the distance image sensors 10A and 10C so that the difference between the first distance image and the third distance image becomes clearer.

  In addition, as long as it has the above-described reflection / transmission characteristics, the object to be irradiated to acquire the third distance image is not limited to light but may be electromagnetic waves having an arbitrary frequency (wavelength).

  For example, when the detection target is water, since there are absorption bands in the microwave, terahertz band, and infrared band, a predetermined wavelength in such a band may be selected as the third wavelength.

  According to the floor surface state detection device of the second embodiment, the floor surface and the liquid can be distinguished and detected, and the liquid existing on the floor surface can be detected in a non-contact manner. The principle will be described with reference to FIG.

  Consider the case where the liquid 200 exists on the flat floor surface 100 as shown in FIG. The first distance image acquired by the distance image sensor 10A is the same as that shown in FIG. That is, an image obtained by integrating the surface shape of the floor and the surface shape of the liquid is acquired as the first distance image (see FIG. 5B). Note that, in FIG. 5B, a portion where the distance image can be acquired among the liquid surface and the floor surface is indicated by a solid line, and a portion where the distance image is not obtained is indicated by a dotted line. The same applies to FIG. 5C described later.

  On the other hand, the light emitted from the distance image sensor 10C is the third wavelength that is absorbed by the liquid and reflected by the floor surface. Therefore, at the position where the liquid exists, the light of the third wavelength is absorbed and does not return to the distance image sensor 10C. The distance image sensor 10C is a TOF method that calculates the distance based on the time until the emitted light is reflected and returned, and therefore the distance becomes infinite when the light does not return. Therefore, as shown in FIG. 5C, the third distance image acquired by the distance image sensor 10C is acquired with an infinite distance for the position where the liquid exists, and the floor for the position where the liquid does not exist. The surface shape is acquired. When the distance is infinite, it is determined as the maximum distance that can be measured by a normal distance image sensor.

  Comparing FIG. 5B and FIG. 5C, FIG. 5B is different from FIG. 3C in that the surface shape of the liquid is acquired. Therefore, when the difference between the first distance image and the third distance image is extracted, the surface shape of the liquid is extracted as shown by the solid line in FIG. FIG. 5 shows a case where the floor surface 100 is flat, but the surface shape of the liquid is similarly extracted even when the floor surface 101 is uneven as shown in FIG. 8 and the liquid 201 exists in the recess. be able to. That is, the surface shape of the liquid can be extracted regardless of the shape of the floor surface.

  As is apparent from the above description of the principle, the floor surface state detection apparatus according to the second embodiment cannot measure the floor surface shape at the position where the liquid exists, and the height of the liquid from the floor surface is unknown. However, if the floor surface shape is measured in advance and stored in the storage unit 12 or estimated from the floor surface shape around the liquid, the floor surface shape at the position where the liquid exists can also be estimated. The height of the liquid from the floor can also be estimated.

  As shown in FIG. 6, the floor surface state detection apparatus according to the third embodiment is obtained by adding contact sensing means 30 to the floor surface state detection apparatus according to the first embodiment.

  The contact sensing means 30 is a means for detecting the presence of the liquid by actually contacting the liquid. For example, a reagent that changes color, volume, weight, etc. by contact with a liquid can be used. The contact sensing means 30 is mounted on a mobile robot or the like. When it is detected that a liquid is present on the floor surface by the method of the first embodiment, the contact sensing means 30 is dispatched to the vicinity of the liquid. Then, the contact sensing means 30 is brought into contact with the liquid to detect whether or not the liquid actually exists in the place, a more detailed range of the liquid, and the like. It is also possible to grasp the slipperiness of the floor surface by collecting the liquid and analyzing the components or measuring the viscosity of the liquid.

  In addition, you may provide the contact sensing means 30 in the floor-surface state detection apparatus of Example 2. FIG. Similarly, more detailed information on the liquid can be detected.

  Also, instead of or in addition to the contact sensing means 30, a high-accuracy camera is installed and dispatched to acquire high-resolution images (distance images and color images), thereby improving the detection accuracy of the liquid position, range, etc. You may plan.

[Variations]
In Examples 1 to 3, the electromagnetic waves are irradiated to acquire the first to third distance images, but the present invention is not limited to the electromagnetic waves, and the distance images are acquired using sound waves. Also good. That is, when both use electromagnetic waves to acquire the first distance image and the second and third distance images (in the case of Examples 1 to 3), the electromagnetic waves are used to acquire the first distance image. When using sound waves to acquire a few distance images, using sound waves to acquire a first distance image and using electromagnetic waves to acquire second and third distance images, the first distance Any of the four patterns in the case of using sound waves to acquire the image and the second and third distance images may be used.

  In the first embodiment, the first and second distance images are acquired by the electromagnetic wave having the first wavelength reflected by the liquid and the second wavelength transmitted through the liquid. In the second embodiment, the first wavelength and the liquid are absorbed. The first and third distance images are acquired by the electromagnetic wave having the third wavelength, and all of the first to third distance images are acquired and the first distance image and the second distance image are obtained. The liquid detection accuracy may be improved by considering both the difference and the difference between the first distance image and the third distance image.

  In the first to third embodiments, not only the distance image sensor but also a two-dimensional color image may be acquired to improve the liquid detection accuracy. Further, the reflection intensity may be measured simultaneously with the distance image, and the distance image may be corrected using the reflection intensity.

  In the first to third embodiments, the TOF method is used as a distance measuring method for acquiring a distance image, but the present invention is not limited to this. Any method can be used as long as it is a distance measuring method using irradiation of electromagnetic waves or sound waves. For example, in addition to the TOF method, a triangulation method, a pattern irradiation method, or the like can be used. Accuracy may be improved by acquiring and correcting a distance image by a plurality of distance measuring methods. However, when acquiring the second distance image, electromagnetic waves or sound waves are transmitted through the liquid, so that the speed of light or the speed of sound changes and an error occurs in distance measurement. Therefore, it is desirable that the distance measurement method has a small distance measurement error due to liquid permeation. Since the triangulation method is a method of measuring the distance according to the reflection angle, the reflection angle changes due to refraction by the liquid, and the distance measurement error becomes large. In this respect, the distance by the TOF method is more liquid than the triangulation method. It is desirable to have less measurement error.

  According to the present invention, since the liquid on the floor surface can be detected in a non-contact manner, various fall prevention measures can be taken accordingly, and the risk of falling can be reduced. For example, monitoring systems for elderly people who assist patients, care recipients, nurses, caregivers, service providers, etc. in hospitals and nursing homes, life support services using mobile robots, surveillance camera systems for business use, automobiles, silver carts, etc. By applying the present invention to an automatic driving support system or the like, a system for reducing the risk of falling and preventing slipping can be realized.

DESCRIPTION OF SYMBOLS 10, 20: Distance image acquisition part 10A-C: Distance image sensor 11: Difference extraction part 12: Memory | storage part 13: Liquid detection part 30: Contact sensing means 100, 101: Floor surface 200, 201: Liquid

Claims (8)

In the floor surface state detection device for detecting liquid existing on the floor surface,
A first distance image acquisition unit that irradiates the floor surface and the liquid with a first electromagnetic wave or a sound wave having a frequency reflected with respect to the liquid and the floor surface, and acquires a first distance image;
A second distance image acquisition unit that irradiates the floor surface and the liquid with a second electromagnetic wave or sound wave having a frequency that transmits the liquid and reflects the floor surface, and acquires a second distance image. When,
A difference extraction unit that extracts a difference between the first distance image acquired by the first distance measurement unit and the second distance image acquired by the second distance measurement unit;
A liquid detection unit that detects the liquid from the difference extracted by the difference extraction unit;
A floor surface state detection apparatus comprising: In the floor surface state detection device for detecting liquid existing on the floor surface,
A first distance image acquisition unit that irradiates the floor surface and the liquid with a first electromagnetic wave or a sound wave having a frequency reflected with respect to the liquid and the floor surface, and acquires a first distance image;
A third distance image obtaining unit that obtains a third distance image by irradiating the floor surface and the liquid with a third electromagnetic wave or sound wave having a frequency that is absorbed by the liquid and reflected by the floor surface. When,
A difference extraction unit that extracts a difference between the first distance image acquired by the first distance measurement unit and the third distance image acquired by the third distance measurement unit;
A liquid detection unit that detects the liquid from the difference extracted by the difference extraction unit;
A floor surface state detection apparatus comprising: A storage unit that stores and holds a plurality of characteristic patterns of the liquid;
The liquid detection unit compares the difference extracted by the difference extraction unit with a plurality of the liquid feature patterns stored in the storage unit to estimate the characteristics of the liquid;
The floor surface state detection apparatus according to claim 1 or 2, wherein   The contact detection means for dispatching to the position of the liquid detected by the liquid detection unit to contact the liquid and detecting the liquid is further provided. The floor surface state detection device according to item. In the floor surface state detection method for detecting liquid existing on the floor surface,
Irradiating the floor surface and the liquid with a first electromagnetic wave or sound wave having a frequency reflected to the liquid and the floor surface to obtain a first distance image;
Irradiating the floor surface and the liquid with a second electromagnetic wave or sound wave having a frequency that transmits the liquid and reflects the floor surface, and obtains a second distance image;
Extracting a difference between the first distance image and the second distance image, and detecting the liquid from the difference;
A floor surface state detecting method characterized by the above. In the floor surface state detection method for detecting liquid existing on the floor surface,
Irradiating the floor surface and the liquid with a first electromagnetic wave or sound wave having a frequency reflected to the liquid and the floor surface to obtain a first distance image;
Irradiating the floor surface and the liquid with a third electromagnetic wave or sound wave having a frequency that is absorbed by the liquid and reflected by the floor surface, and obtains a third distance image;
Extracting the difference between the first distance image and the third distance image, and detecting the liquid from the difference;
A floor surface state detecting method characterized by the above. Comparing the difference with a plurality of stored feature patterns of the liquid to estimate the feature of the liquid;
The floor surface state detection method according to claim 5 or 6,   8. The contact sensing means for detecting the liquid by contacting the liquid is dispatched to the position of the liquid detected by the difference, according to claim 5. Floor surface detection method.

Images